@MelvinInvests

AI Analyst @MilkRoadAI | Finding opportunities across AI, photonics, defense, space, and tech.

$1 to see my portfolio →
Joined June 2026
Elon Musk just revealed the biggest threat to America’s AI dominance. It is not chips or software but rather electricity. Here are five under the radar stocks positioned to WIN (Save this).
7
6
34
5,914
Melvin retweeted
Elon Musk just revealed the biggest threat to America’s AI dominance. It is not chips or software but rather electricity. Here are five under the radar stocks positioned to WIN (Save this).
7
6
34
5,914
$POWL builds switchgear and electrical-control systems for large power users. Its equipment safely distributes electricity across data centers, utilities and industrial facilities. The company has already received a data-center order worth more than $400 million.
2
10
699
$PLPC makes connectors, protective hardware and other components used across power lines and substations. These products are small but essential whenever utilities build or reinforce transmission networks. Its energy sales have already accelerated alongside transmission demand.
1
7
752
Before you read on, if you want to know exactly which of these stocks I’m putting my own money into, you can track my full portfolio here. You’ll also get constant updates whenever I buy, add, trim, or sell a position. link.milkroad.com/tx5qip
1
6
466
$NVT supplies electrical connections, enclosures and cooling infrastructure. Its equipment protects power systems while helping increasingly dense server racks manage heat. It can benefit from both the electricity and cooling bottlenecks created by AI.
1
7
507
$ATKR manufactures electrical conduit, cable-management systems and supporting infrastructure. Its products protect and route the wiring carrying electricity through data centers and industrial facilities. More facilities and greater power density can increase the electrical content required per building.
1
6
581
$AMSC supplies equipment that stabilizes voltage and improves power quality. That becomes increasingly important as massive AI campuses place more volatile loads on the grid. Its grid business already generates most of its revenue.
2
8
608
Melvin retweeted
Trump just gave America’s biggest AI companies room to keep racing, if they agree to police themselves. President Trump brought the leaders of major AI companies to the White House, where they signed a voluntary agreement for developing superintelligence more safely. Executives from companies including Meta, OpenAI, Nvidia, Google, Microsoft, Anthropic, Amazon, xAI and Palantir attended the meeting. Under the agreement, companies will build stronger internal controls, conduct their own risk reviews, bring in outside auditors and evaluators, and require their boards to independently review the auditors’ reports. Zuckerberg described it as an industry starting point rather than a complete safety framework. The key point is that this is self regulation, not a strict new federal law. Trump is allowing AI companies to continue moving quickly while asking them to monitor their technology and document their safety work. The DOJ and FBI would remain available for serious legal or national security emergencies rather than supervising development every day. For AI companies, this is largely positive because it reduces the immediate threat of strict licensing requirements, development limits or a government mandated slowdown. Meta, OpenAI and their competitors can continue investing aggressively while telling customers and lawmakers that their systems face independent review. For the market, the message is supportive of the broader AI trade. The Trump administration appears to view superintelligence as a strategic race with China that the United States must win and that supports continued spending on AI models, agents, chips, data centers, networking equipment and power infrastructure. This is overall a win for the AI industry!
6
5
30
3,257
There is a lot of good news coming out of OpenAI today and it is helping the AI trade stay strong even as higher Treasury yields pressure the broader market. ChatGPT is now at 1.2 billion weekly users, roughly 2.5 million businesses are using OpenAI and the company has launched more than 40 models as it continues expanding across consumer and enterprise AI. At the same time, OpenAI announced Dots, its response to Meta’s Muse, while its annualized revenue run rate is reportedly approaching $70 billion. Enterprise sales have more than doubled since July. OpenAI is also reportedly looking to raise at least $30 billion at a valuation of roughly $1.4 trillion, with some reports suggesting it could target as much as $1.5 trillion. That would give the company even more capital to keep spending aggressively on compute and infrastructure without needing to immediately go public. The larger story is the growing race to control the AI agent market. These agents will eventually become the main way people search, shop, communicate, advertise, write software and complete everyday tasks. OpenAI has the stronger standalone AI brand but Meta can distribute its agent through Facebook, Instagram, WhatsApp and its advertising network. That gives Meta a major advantage. It can offer its agent cheaply or for free, improve it using data from billions of users and monetize it through advertising. As an agent learns someone’s preferences, contacts and routines, switching to another service could also become more difficult. The agents themselves may eventually become commoditized as OpenAI, Meta, Google, Anthropic and xAI copy one another’s best features. The real advantage could come from distribution, user data, memory, application integrations, available computing power and trust. This race will also require a massive infrastructure buildout. AI agents use more computing power than normal chatbot requests because they can plan tasks, search databases, operate applications and check their own work. The International Energy Agency expects the spread of AI and agent based applications to help double global data center electricity use by 2030. The market is looking past higher yields because investors see two major opportunities here, AI agents becoming the next major software platform and the enormous infrastructure buildout needed to support them. If you want to see how I’m positioned around the AI agent race and the infrastructure needed to support it, check out my full Milk Road Pro portfolio below. link.milkroad.com/d8cj6v
5
13
2,119
For the benefit of humanity, I think you should stop shorting stocks.
For the benefit of humanity, the markets should tank hard and prevent the OpenAI and Anthropic IPOs.
3
2
34
4,140
HOLY SMOKES! SpaceX may have just landed one of the largest AI infrastructure agreements ever reported. Anthropic has reportedly agreed to pay SpaceX up to $84.5 billion for access to Nvidia based computing capacity through 2029 and that is nearly double the roughly $45 billion maximum value SpaceX previously disclosed. SpaceX is building an AI cloud business that could eventually compete with CoreWeave, Nebius, Microsoft, Amazon, Google, and Oracle. ARK models SpaceX’s active AI compute capacity increasing from 0.3 gigawatts in 2024 to 66.6 gigawatts by 2031. Terrestrial data centers drive the early growth, but ARK expects orbital compute to become a major source of capacity after 2029 and potentially represent more than half of the total by 2031. The strategy would begin with SpaceX building data centers on Earth for customers such as Anthropic like @elonmusk has already done and then SpaceX could then use Starship to launch computing systems, solar arrays, networking equipment, and radiators into orbit. Starlink would provide the communications network connecting those orbital systems with customers and ground infrastructure and this creates a level of vertical integration that no traditional cloud company can easily match. SpaceX could control the launch vehicle, satellite network, communications layer, terrestrial data centers, orbital computing systems, and customer contracts inside one company. Orbital computing will also avoid some of the grid connection delays, land shortages, and permitting problems slowing data center construction on Earth. I’m already a proud SpaceX investor, and if you want to see the other AI infrastructure and space names I’m positioned in around this buildout, check out my portfolio below. link.milkroad.com/tx5qip
19
52
7
405
34,313
I’m an analyst at Milk Road, and my job is to find underrated gems before the market catches on. We called names like MU, CRDO, NBIS, and BE over the last 6 months. Join me and my team for just $1. link.milkroad.com/tx5qip
1
801
Melvin retweeted
HOLY SMOKES! SpaceX may have just landed one of the largest AI infrastructure agreements ever reported. Anthropic has reportedly agreed to pay SpaceX up to $84.5 billion for access to Nvidia based computing capacity through 2029 and that is nearly double the roughly $45 billion maximum value SpaceX previously disclosed. SpaceX is building an AI cloud business that could eventually compete with CoreWeave, Nebius, Microsoft, Amazon, Google, and Oracle. ARK models SpaceX’s active AI compute capacity increasing from 0.3 gigawatts in 2024 to 66.6 gigawatts by 2031. Terrestrial data centers drive the early growth, but ARK expects orbital compute to become a major source of capacity after 2029 and potentially represent more than half of the total by 2031. The strategy would begin with SpaceX building data centers on Earth for customers such as Anthropic like @elonmusk has already done and then SpaceX could then use Starship to launch computing systems, solar arrays, networking equipment, and radiators into orbit. Starlink would provide the communications network connecting those orbital systems with customers and ground infrastructure and this creates a level of vertical integration that no traditional cloud company can easily match. SpaceX could control the launch vehicle, satellite network, communications layer, terrestrial data centers, orbital computing systems, and customer contracts inside one company. Orbital computing will also avoid some of the grid connection delays, land shortages, and permitting problems slowing data center construction on Earth. I’m already a proud SpaceX investor, and if you want to see the other AI infrastructure and space names I’m positioned in around this buildout, check out my portfolio below. link.milkroad.com/tx5qip
19
52
7
405
34,313
Melvin retweeted
Bloom Energy is up more than 11% today, and this chart explains exactly why (Save this). Global data center electricity demand is projected to rise from roughly 400 terawatt hours in 2024 to 1,570 terawatt-hours by 2032 and that would represent nearly fourfold growth in only eight years. Utilities often need years to approve grid connections, expand transmission networks, and build enough generation capacity to support a large AI data center. AI companies cannot afford to wait that long because every delayed data center leaves expensive GPUs sitting idle and prevents the operator from generating cloud revenue. Bloom Energy offers an alternative by installing fuel-cell systems that generate electricity directly at the data center. Bloom says it can deploy these systems in approximately 90 days, while traditional utility upgrades can take several years. This speed makes Bloom more than a clean-energy company because it effectively makes Bloom a time-to-power company. A more expensive megawatt available today can be worth more than a cheaper megawatt that arrives several years later. Bloom is not the only company positioned to benefit from this trend. GE Vernova can supply the natural gas turbines and grid equipment required to power enormous AI campuses, while Caterpillar and Cummins can provide rapidly deployable generators and distributed power systems. Generac could benefit as data centers adopt more microgrids, backup generation, and on site energy management systems. Eaton, Powell Industries, and nVent do not generate electricity, but they provide the switchgear, power distribution, electrical connections, and control equipment required to move that power safely through a data center. Vertiv benefits after the electricity reaches the facility because higher rack densities require increasingly advanced power management and liquid cooling. Fluence could benefit from the need for large battery systems that stabilize on-site generation, manage peak demand, and provide backup power. Oklo and NuScale offer longer-term exposure through small nuclear reactors that could eventually provide continuous power directly to large data-center campuses, although those projects face much longer construction and regulatory timelines. The full value chain includes the companies that generate, store, distribute, manage, and cool every additional megawatt consumed by AI. If you want to see the other power and AI infrastructure names I’m positioned in alongside Bloom Energy, check out my full Milk Road Pro portfolio below. link.milkroad.com/d8cj6v
4
4
20
3,281
OpenAI is approaching $70 billion in annual recurring revenue and the biggest winner will be Oracle (Save this). OpenAI’s enterprise business is now its fastest growing segment, with B2B revenue reportedly more than doubling since July. This matters because businesses generate recurring inference demand every time employees use ChatGPT, Codex, or AI agents. Oracle has a partnership exceeding $300 billion with OpenAI to provide huge amounts of computing power. The biggest concern was whether OpenAI could grow fast enough to pay for all the infrastructure it committed to, and this revenue growth makes that risk look much smaller. But even if OpenAI eventually loses its lead, Oracle’s infrastructure will still capitalize. AI probably will not be a winner take all market. Different models will be better at coding, drug discovery, business automation, and other tasks, but they will all need roughly the same GPUs, data centers, power, and networking and that means Oracle could shift capacity toward other AI labs and enterprise customers if necessary. Oracle is not only betting on OpenAI but rather that demand for AI computing will keep growing across the entire industry. With Oracle’s backlog reaching $664 billion and its cloud infrastructure revenue growing 121% last quarter, OpenAI’s growth makes the bull case stronger but Oracle does not need OpenAI to be the only winner for its AI bet to work. If you want to see the other AI infrastructure names I’m positioned in alongside Oracle, check out my full Milk Road Pro portfolio below. link.milkroad.com/tx5qip
OpenAI just shipped the best model in the world and the biggest winner will be Oracle (Save this). GPT 6 Astra launched yesterday and OpenAI calls it the most intelligent and aligned model in the world and it scored 97.6% on FrontierMath Tier 4, 98.6% on ARC-AGI-3 against just 7.8% for GPT-5.6 Sol, 95.9% on BenchCAD Vision2Code, 72.6% on OSWorld computer use and 100% on ExploitBench. OpenAI positions it as a computer use model that operates browsers, spreadsheets, CRM systems, terminals, and desktop applications the way a person does, finishing multi-step jobs rather than describing them. It completes OSWorld tasks 47% faster than its predecessor, cutting average task time from about 75 minutes to 40 and that distinction matters enormously for compute demand. A chatbot answers a question and stops while an agent that operates software runs for forty minutes, burns tokens continuously, and gets deployed thousands of times in parallel across an enterprise. Astra also carries a 1.05 million token context window and costs $10 per million input tokens and $50 per million output, roughly 2.5 times Sol's launch pricing. Higher token consumption at higher prices per token is the single most bullish combination possible for whoever rents OpenAI its compute. OpenAI's business is already compounding fast because its revenue run rate topped $40 billion in August, roughly double the $20 billion it ended 2025 with, enterprise is now more than half of revenue, and the ads business alone hit a $1 billion annualized run rate in about 200 days which brings us to Oracle, because OpenAI is their biggest customer. Now look at Oracle's remaining performance obligation which is contracted revenue Oracle has signed but not yet recognized, so it is the closest thing to a forward order book that exists in software. At $638 billion it is up 363% year over year and grew $85 billion sequentially in a single quarter. The most underappreciated part is how these contracts are structured. Oracle disclosed that most of the RPO increase in Q3 and Q4 came from large AI contracts where customers either prepaid for GPUs or bought and supplied the GPUs themselves, and the prepaid plus customer-supplied hardware portion now totals $75 billion. That reframes the entire bear case. The loudest criticism of Oracle is that it is borrowing enormous sums to build data centers for one financially unproven customer but $75 billion of the hardware is already funded by customers, which materially reduces the capital Oracle must raise itself. Management expects 12% of the RPO to convert to revenue within twelve months and another 34% between thirteen and thirty six months, and both conversion rates are accelerating. Against a $638 billion backlog, that is roughly $77 billion recognizable in year one and another $217 billion in the following two years.The reflexive loop is what makes this compelling. Better models drive more agent deployment, agents consume far more tokens than chat, more tokens require more compute, more compute flows to Oracle under a contract that does not even begin until 2027, and Oracle's OpenAI revenue is therefore levered directly to OpenAI's product quality rather than to Oracle's own sales execution. Astra is evidence that quality is improving and OpenAI is diversifying its funding base toward an IPO, which reduces the risk that its compute commitments outrun its ability to pay. If you enjoyed reading this, make sure to follow @MelvinInvests for more AI infrastructure and semiconductor insights and turn on post notifications so you don't miss a single update. If you want to see exactly what I'm buying as an analyst at Milk Road Pro, check out the link below. link.milkroad.com/tx5qip
10
3
24
7,102
If you own Micron stock, You NEED to read this! Wall Street is currently expecting roughly $51 billion in revenue and around $31.50 to $31.60 in adjusted EPS. I think Micron comes in higher. My estimate is around $52.2 billion in revenue, 87.5% gross margins and $33.00 in adjusted EPS. That would put my revenue estimate roughly $1.2 billion above the Street and EPS around 4% to 5% above consensus. The reason I am comfortable being slightly above the Street is pretty simple. Memory pricing continues to remain extremely strong while supply is still struggling to keep up with demand. Micron originally guided for around $50 billion in revenue, 86% gross margins and $31.00 in EPS, so expectations have already moved meaningfully above management's original outlook. I think pricing has remained strong enough for Micron to come in above those expectations again. Gross margin is going to be one of the most important numbers for me. The Street is already looking for margins around 87%, which is insane when you think about where this business was just a few years ago. I am looking for around 87.5%. If Micron can continue growing revenue while keeping margins anywhere near these levels, the amount of earnings and free cash flow this company can produce becomes ridiculous. DRAM is another big reason I am staying above consensus. AI infrastructure continues to become more memory intensive with every generation. Larger models, longer context windows, inference and eventually AI agents all require more memory alongside the compute. At the same time, customers are trying to lock up supply years in advance because they are worried future capacity will not be available when they need it. This is also why the long term agreements are probably one of the most important things I will be listening for on the call. Last quarter Micron had signed 16 strategic customer agreements and I want to know how much that number has increased since then. Are we talking about 18 agreements now? 20? Even more? More importantly, I want to know whether the size and duration of these agreements are getting larger as customers become increasingly worried about securing enough memory. These agreements matter because they give Micron something the memory industry historically has not had much of, which is long term demand visibility. Memory has always been extremely cyclical because manufacturers usually do not know exactly what demand or pricing will look like several years into the future. If customers are now willing to commit to supply years in advance, provide deposits, agree to minimum purchases or accept pricing protections, Micron suddenly has much better visibility into future revenue and utilization. I also want more detail around HBM because this is becoming an increasingly important part of the Micron thesis. The company has already started high volume HBM4 shipments for its lead customer and has sent qualification samples to additional customers. I want to know whether Micron is maintaining or gaining HBM market share as the industry moves from HBM3E into HBM4 and eventually HBM4E. If Micron can hold around 20% share or move even higher, it becomes an even bigger beneficiary of the AI infrastructure buildout. HBM4E is another area I want management to talk about. I want to know when additional customers are expected to qualify the product, how much of 2027 HBM capacity is already committed and whether customization allows Micron to capture even better pricing. The more customized HBM becomes for specific AI accelerators, the harder it becomes to look at Micron as just another commodity memory company. The most important part of the entire earnings report, though, is probably going to be next quarter guidance. The market already expects a monster fiscal Q4, so whether Micron reports $51.8 billion, $52 billion or $52.5 billion might not matter nearly as much as what management says comes next. I would personally like to see Micron guide toward roughly $58 billion in revenue next quarter, gross margins around 88% and EPS somewhere in the high $30s. If management gets anywhere near those numbers, that would tell me the memory pricing environment is still extremely strong heading into 2027. The biggest risk I want addressed is supply. Samsung, SK Hynix, Micron and Chinese memory manufacturers are all investing heavily because the economics are so attractive right now. Eventually that capacity will come online (I personally believe it has no meaningful effect but I want the management to address the Chinese memory makers) .The real question is whether AI memory demand can continue growing faster than the industry can add supply. I want management to give us more clarity on 2027 DRAM supply growth, how quickly new fabs can actually contribute meaningful capacity and whether Chinese memory companies are starting to change the supply picture. If supply suddenly starts growing much faster than demand, pricing could turn very quickly, but right now I still think demand is winning. I also want to hear more about capital returns. Micron is spending aggressively because it needs more capacity but if revenue and margins remain anywhere near current levels, the company should also generate an enormous amount of cash. At some point investors are going to start asking how much of that cash gets reinvested versus eventually being returned through buybacks or other capital returns. Overall, I am going into Wednesday expecting another beat. My numbers are around $52.2 billion in revenue, 87.5% gross margins and $33.00 in adjusted EPS, compared with the Street at roughly $51 billion and around $31.50 to $31.60 in EPS. But I am not going to make a prediction on whether the stock goes up or down after earnings because honestly that is basically a coin toss at this point. Micron has already had a massive run and expectations are extremely high. The company could beat the Street and still sell off if guidance does not clear the bar investors have built into the stock. It could also report numbers close to expectations and move higher if management gives an extremely bullish outlook for pricing, long term agreements and 2027 demand. That is why I care much more about what Micron actually tells us about the business than trying to guess how the stock trades the next morning. If you enjoyed reading this and you want to see exactly how I’m positioned in Micron ahead of earnings and the rest of the memory names I hold, check out my portfolio below. link.milkroad.com/tx5qip
3
6
35
4,730
Shorting the US dollar makes a lot of sense here. Higher Treasury yields normally strengthen the dollar because they make US assets more attractive but this time, yields are not rising because the economy is suddenly stronger. They are rising because oil is driving inflation higher, the government is issuing enormous amounts of debt and investors want more compensation to hold it. US debt has already crossed $40 trillion, while annual interest costs now exceed $1 trillion. As yields rise, refinancing that debt becomes even more expensive, forcing the government to borrow more just to cover its growing interest bill. Ray Dalio explains the bigger problem well because bonds are promises to deliver dollars in the future. If investors begin to believe those future dollars will be worth less, they will demand higher yields or move into alternative stores of wealth such as gold. China’s actions appear to reflect that shift. Its reported Treasury holdings fell to $618 billion in July, their lowest level since 2008 and less than half the peak of roughly $1.32 trillion reached in 2013. At the same time, China has been accumulating gold. The People’s Bank of China increased its reserves for a 22nd consecutive month in August, bringing its holdings to a record 76.73 million ounces. That does not mean China is suddenly abandoning the dollar but it does show that one of America’s largest foreign creditors has been gradually reducing its exposure to US debt while increasing its exposure to an asset that cannot be printed. That puts policymakers in a difficult position. They can leave rates extremely high and risk damaging the economy or the Federal Reserve can eventually step in to support the bond market by creating more money. That could bring yields down but it would also weaken the dollar and reduce its purchasing power. The oil shock only accelerates this problem. Higher oil keeps inflation elevated, makes rate cuts harder and pushes the government’s borrowing costs even higher. The dollar may continue benefiting in the short term but the longer this lasts, the more pressure builds beneath the surface. And as Dalio points out, investors do not necessarily need to replace the dollar with another currency. They can move into gold, commodities, stocks and other scarce assets instead. I’m already positioning around this weaker dollar thesis. If you want to see exactly what I hold, check out my full my portfolio below. link.milkroad.com/tx5qip
16
26
81
24,250
I’m an analyst at Milk Road, and my job is to find underrated gems before the market catches on. We called names like MU, CRDO, NBIS, and BE over the last 3 months. Join me and my team for just $1.milkroad.com/pro/?utm_medium…
1
1,030
Why was there a sell off today despite all the good news? Nvidia announced another $150 billion in share buybacks, Meta launched a new enterprise AI platform and hired MongoDB CEO CJ Desai to lead it, and SpaceX put Starship into orbit while deploying 26 Starlink satellites. These are exactly the kinds of news that should push the market higher yet it sold off. The market clearly wants to move higher and Nvidia gained while the broader market fell, and the Nasdaq entered the day near record highs after another strong week for AI stocks. Investors are still willing to buy companies delivering strong earnings, rising AI demand, and positive news. Honestly, I am surprised by how well the market is holding up. The 10 year Treasury yield has climbed above 5.2% to its highest level since 2007, yet the S&P 500 remains close to its record and the Nasdaq just finished another positive week. That resilience shows how much demand is still sitting underneath the market. The problem is that yields are becoming increasingly difficult to ignore. When government bonds pay more than 5%, investors can earn a strong return without taking the same risk as owning stocks. That makes expensive technology companies less attractive and lowers the valuation investors are willing to pay for their future earnings. Higher yields also make mortgages, business loans, and corporate debt more expensive. That can slow consumer spending, delay new projects, raise the cost of building data centers, and eventually weaken company earnings. The market is now caught between two powerful forces. Strong earnings, AI spending, and positive company news are pulling stocks higher, while the 5.2% Treasury yield is pulling valuations lower. I am still very bullish over the long term because the AI investment cycle and earnings story remain strong. But over the short term, the market is honestly a coin toss. If yields stabilize or move lower, stocks could quickly break higher. If yields continue climbing, even the best company news may not be enough to stop another pullback.
7
8
2
77
9,697
CPU demand is EXPLODING because of AI agents like Muse and Grok. The first wave of AI spending focused on GPUs because they provide the parallel computing required to train and run large models but agents need more than intelligence. An agent may require a persistent cloud based virtual machine to manage memory, permissions, credentials, files, applications, and network activity. The GPU performs the model inference, while the CPU runs the surrounding operating environment that allows the agent to complete real work. Meta’s Muse provides an early example and as Meta, xAI, OpenAI, and other companies scale similar agents, millions of users could eventually require continuously available computing environments. The supply chain is already showing signs of pressure. TrendForce now classifies server CPUs as tight, with current lead times of 25 to 30 weeks compared with a balanced range of 16 to 20 weeks. That means customers are waiting approximately two months longer than they would in a normally supplied market. The rest of the server supply chain is tightening as well. DRAM and hard drives are classified as very tight, while enterprise SSDs, ABF substrates, and MLCC components are also experiencing constrained supply. This matters because AI agents do not consume CPUs in isolation. Every CPU server also requires memory, storage, networking, packaging substrates, power components, and data center capacity. JPMorgan believes this change will reshape the composition of AI servers. The bank estimates that the GPU to CPU ratio could decline from four GPUs per CPU in 2024 to just 1.9 GPUs per CPU by 2028. Data centers will move from buying one CPU for every four GPUs to buying more than one CPU for every two GPUs. If total GPU deployments continue rising while the CPUs required per GPU nearly double, CPU demand could grow much faster than the market expects. If you want to see the companies I hold to capture that shift, check out my full Milk Road Pro portfolio below. link.milkroad.com/tx5qip
Meta’s Muse agents are about to ignite the next CPU boom (Save this). The more tasks these agents complete, the more CPU power Meta will need to run them. Here are the five stocks positioned to win from this.
7
14
1
82
11,865