Serenity, an anonymous AI and semiconductor supply-chain analyst with more than 840,000 followers, published a detailed post on June 16, 2026 arguing that the AI bubble will not burst this year. His central call was blunt: spending on AI infrastructure is still in expansion mode, and the capex cycle could continue through 2028.
In his view, AI is a transformative technology on the scale of the agricultural and industrial revolutions. He argued that if the industry reaches superintelligence, the economic upside could be difficult to quantify under current models, touching areas such as disease treatment, faster quantum computing, and stronger corporate profitability.
U.S. strategy and hyperscaler cash flow form the core thesis
Serenity pushed back on bubble concerns from two angles. First, he said the U.S. has a strong incentive to keep funding AI infrastructure in order to preserve an edge over China in areas including military capability and cybersecurity. That means support at the national level is unlikely to disappear even if large language model training and inference have not yet produced broad-based profitability.
Second, he pointed to the balance sheets of major cloud providers. Microsoft, Google, and Amazon all have substantial free cash flow that can support heavy capital expenditure. He described Oracle as more cautious and Meta as neutral, yet still argued that the broader group of tech giants has enough financial capacity to keep spending. He also said the Federal Reserve is unlikely to deliver an aggressive tightening shock in the near term.
Debt-sensitive infrastructure names face risk, but upstream chips still stand out
Serenity did not present the sector as risk-free. He specifically warned that some infrastructure companies relying heavily on debt could be exposed to bubble risk, naming CoreWeave as an example. He also flagged circular valuation dynamics tied to OpenAI pre-booking agreements and GPU purchase arrangements between Nvidia and newer cloud providers.
At the same time, he said recent large-scale fundraising by OpenAI has eased some of the broader systemic concern. His strongest conviction remains in the upstream hardware chain. He said semiconductor names ranging from LITE to SK Hynix still show meaningful real profit potential and do not appear bubble-like to him. Based on guidance from companies such as Broadcom, he expects AI capex to accelerate into 2028, with many firms already operating under multi-year contracts. In his phrasing, the music is at least not stopping this year.
Analyst gained attention through AI hardware bottleneck research
Serenity’s following has been built in part on his “chokepoint” framework for the AI hardware supply chain. He describes himself as a former AI research scientist and a member of the RISC-V Foundation. Before building an audience on X, he was active on Reddit’s WallStreetBets, then shifted his focus to identifying irreplaceable bottlenecks in areas such as photonics, advanced packaging, and specialty semiconductor materials.
The source article said his portfolio was rumored to have returned 225x in two years, and that his research has been cited by Bloomberg and Reuters. Despite the attention from the market and reported interest from hedge funds, Serenity has remained anonymous and does not give interviews.