The AI bear case has evolved into a comprehensive argument covering three distinct areas, according to recent analysis. Michael Burry's accusations of earnings manipulation through extended asset depreciation schedules have gained attention, with his calculations showing $176 billion of understated depreciation across the industry from 2026 through 2028. However, as reported by Investing.com India, the case also includes circular financing concerns involving Nvidia-OpenAI-Oracle relationships and revenue generation questions based on MIT's GenAI Divide study showing 95% of enterprise AI pilots showed no measurable profit impact. The analysis reveals that while two of these arguments present legitimate concerns, one falls apart when examined closely. The problem with the way the AI bear case is usually discussed is that skeptics blur three separate claims into a single mood, each pointing to something real but often getting stretched beyond what the evidence supports.
Burry's primary argument centers on hyperscalers depreciating Nvidia hardware over five to six years when the real economic life is closer to two to three years. According to the analysis, this creates a $176 billion gap in understated depreciation across the industry through 2028. The report notes that Oracle overstated earnings by nearly 27% and Meta by 21% by 2028 using similar accounting methods. However, the analysis points out that depreciation is a non-cash charge that doesn't affect operating cash flow, with companies like Alphabet still generating approximately $165 billion in operating cash flow in 2025. The argument focuses on earnings quality rather than fundamental business performance, with the key distinction being that you can be right about the fundamentals and still pay far too much for them. As noted by recent analysis, Cisco had real revenue in 2000 but still fell 80% and took the better part of two decades to reclaim that high.
The second bear argument addresses circular financing arrangements totaling over $800 billion in vendor relationships. As reported by Investing.com India, UBS estimates the OpenAI-Nvidia arrangement represents up to 13% of Nvidia's projected 2026 revenue, while the other 87% comes from traditional customer relationships. The analysis acknowledges legitimate concerns about OpenAI reportedly losing around $14 billion this year and vendor-warrant deals like AMD's engineering. However, the report emphasizes that most hyperscaler AI revenue comes from enterprises and consumers paying real money rather than circular financing loops, with the circular piece representing a caution flag rather than a bubble thesis. The concern is legitimate for specific deals, but the mistake is treating the loop as the whole story. UBS put the OpenAI-Nvidia arrangement at up to 13% of Nvidia's projected 2026 revenue, while the other 87% comes from customers buying at arm's length.
The third argument, based on MIT's GenAI Divide study showing 95% of enterprise AI pilots showed no measurable profit impact, faces significant data challenges. According to the analysis, the failures were organizational, not technological, with companies building in-house tools instead of purchasing them and targeting marketing rather than back-office operations. The report notes that roughly 90% of workers reported using personal AI tools at work against only 40% of firms with official subscriptions. Most critically, the analysis reveals that Morningstar estimates the U.S. AI sector produced around $100 billion in services revenue in 2025, enough to cover model training costs. The key misunderstanding is that the study measures the buyer's return, not the seller's revenue. The question was never whether the revenue is real, but whether that revenue eventually covers model training and research, not just the cost of inference. The 'no revenue' narrative is largely noise for diversified investors, as the sector's revenue generation capabilities are being misrepresented in the bear case.
Despite legitimate bear arguments, the analysis identifies free cash flow as the primary concern rather than revenue generation. The report notes that the four hyperscalers spent approximately $410 billion on capex in 2025, with 2026 guidance pointing to around $700 billion. Pivotal Research projects Alphabet's free cash flow could drop nearly 90% this year to roughly $8 billion from $73 billion, while Amazon's free cash flow could turn negative. However, the analysis distinguishes between companies spending to build the next generation versus those bleeding cash to prop up dying models. There's a difference between a company bleeding cash to prop up a dying model and one spending record sums to build the next one. Alphabet, Amazon, and Microsoft aren't buying back stock at these levels - they're plowing the cash back into the business. The honest AI bear case isn't about earnings manipulation or lack of demand, but about spending running years ahead of payback, and every hyperscaler building as if the return has already been proven.
