InvestAnswers: AI bubble at "1998 stage," DeFi Dev leaps warn of 130% dilution trap

Where the AI trade stands versus the dot-com bubble

James compares today's AI run to the dot-com bubble he traded through between 1995 and 2000. He says the 1995 dot-com move "gained 985% in 62 months," while the current AI cycle is "about 47 months through" and up roughly 157%, with the chip sector up 345% since November 2022. He recalls shorting EMC in late 1999 after valuations stopped making sense, and notes the Nasdaq PE reached 60x at the March 2000 peak versus roughly 20x today.

He argues current AI valuations are far from bubble territory: "Nvidia today is at 14x forward earnings. Micron is at 5.9 times today's earnings... they're stupid cheap levels." He also cites "another $10.3 trillion dollars of AI buildout to come."

His own read is that the cycle sits at "circa first half of 1998" on a dot-com-mapped chart, implying roughly 24 months left before a parallel top, with the IA13 list built with targets running "up to the end of 2028." He flags risks to watch — "debt," "defaults," "slowing growth" — and mentions an Anthropic growth chart he shared on Patreon showing the company "hasn't grown since July" while reportedly seeking a "$2 trillion IPO."

Ranking the IA13 names and portfolio approach

James ranks his IA13 picks by upside: top tier includes Broadcom, Arm, and Marvell; mid-tier includes Tesla, Nvidia, and Micron; bottom tier (not necessarily bad, just less upside) includes Palantir, Google, and TSM. He notes Palantir traded near "210" after being at "106 12 weeks ago."

Responding to a question about building positions with a "low six figure cash" amount, he points to his "Retire on Target" share counts — for example 655 shares of SpaceX and 322 of Tesla — and urges building "multiple retire on bags," buying dips like Nvidia's fall to "$88 during the tariff tantrum" or Palantir's drop to 106.

The dangers of proxy wrappers: PE1 and smarter web

On a listener's SpaceX exposure via the ASX-listed PE1/Pangaea fund, James is blunt: the fund claims "over 550 holdings," which he calls "complete nonsense," amounting to only "4% real SpaceX exposure." He notes the wrapper is "down 29.5% from pre-IPO peak" while "the SpaceX reality is up 24.4%," and warns against locking money into funds with "2 and 20%" fee structures and multi-year lockups.

On Smarter Web Company versus MicroStrategy, he compares NAV premiums — Smarter Web at 1.37 versus MicroStrategy's 0.74 — and notes Smarter Web carries "$32 million in debt" and just "$2 million in cash," while MicroStrategy "hold 309 times more Bitcoin" and far greater liquidity. He says there's "no comparison side by side" between the two.

DeFi Dev Corp leaps and Solana's burn mechanics

On DeFi Dev Corp as a Solana proxy, James warns that despite a NAV premium of "0.51," the stock carries "nearly half the market cap" in debt and has diluted shares by "130.76%" since June 2025 — "5 and a half times higher than that of MicroStrategy." He notes Solana has "beaten DeFi Dev by 251%" since both started stacking, and that Cantor Fitzgerald's target on the stock fell from "$45" to "$10."

"Think debt and dilution, not death and taxes, because they will kill you on the proxy side of the world." — James

On a separate Solana question, James breaks down the burn mechanic: "5,000 lamports per signature," half burned, with about "60,000 minted daily" against "434 and 990 burned daily" currently. He calculates that at 10 billion transactions a day over a year, "1.44% of the supply will be burned," rising to "3.59%" at 25 billion transactions a day, which he calls bullish for supply even as validators keep priority fees.