Decentralized AI and on-chain derivatives are both heading in the direction I expected, and the two coins I think benefit most just got a lot cheaper. That's why I'm buying TAO and HYPE now, sized so I can add if they fall further.
The answer in brief
I'm buying Bittensor and Hyperliquid. Both are long-term bets, and both are cheap after a few horrible weeks.
Bittensor because I think the future of AI is open source and decentralized, and the usage data now backs that up.
Hyperliquid because I think tokenized assets and on-chain derivatives take it to hundreds of dollars, and it's around $65 as I write this.
We're in Q4 of the bear market, and that's good enough for me to act now, even though it isn't perfectly timed.
If the open source trend reverses, or if Hyperliquid's revenue and burns stop growing, the bets don't work.
Why I'm buying now
Horrible week, nothing else to say.
Kinda feels like "it's not even our fault" too.
Why does BTC and alts get dragged down with cheating CEOs and whatnot?
Well, there's a few things happening in the space that concern me.
... but I think it concerns the market even more!
So today, I've made some big moves.
Not because this is PERFECTLY timed, or all the Fibonacci levels align with the moon and stars...
... It's mostly because we're in Q4 of the bear market, we've had a few horrible weeks, and the long-term convictions I'm looking to buy are cheap.
Let me show you what I mean by cheap.
The cycle keeps on cycling..
And BTC is yet again down into its bear bottom buy zone. That's the chart below.
The orange box is the 2021 high. BTC came back down to it earlier this year for the 2026 low, and now it's back in the box again.
As I write this, BTC is around $64,300. That's down about 21% in the past 30 days, and roughly 30% from the September top if you measure to this week's low.
It might fall further. The 2022 low went slightly through the 2017 high before it turned, so I wouldn't be shocked if this one does the same.
But this is "good enough" for me.
And here's the point for the two coins in this article. If BTC is in its buy zone, the alts I actually believe in have been hit harder. TAO is down almost 50% in a couple of months. HYPE is down 40%.
I expect BTC to reach somewhere between $150K and $200K over the next 3 years. That's roughly a 3X from here.
Both bets below need to beat that to be worth the extra risk. I think they can, and I'll explain why below.

Bet one: Bittensor, and the flip to open source
Decentralized AI is the future in my opinion.
And the data backs it:
open source (open weights) models are now used MORE THAN CLOSED MODELS.
That's right, in 2026 the big guys (OpenAI, Anthropic, Google) have lost SIGNIFICANT market share.
The two charts below show usage share on OpenRouter. Blue is open weight models, orange is the big AI companies.
The first chart is one year ago. On November 17, 2025, the closed models had 74.4% of the usage. Open weights had 25.6%.
The second chart is today. On November 9, 2026, open weights had 75.3%. The big guys only have 24.7%.
It's completely flipped in 12 months.
Look at how it happened too. It didn't happen in one jump. Open weights crossed 50% around February, dipped back in March, and then climbed steadily from May onwards. It's been above 65% since June.
So why does this matter for Bittensor?
Bittensor is a bet that AI gets built and served by an open, decentralized network instead of three or four companies. If businesses and agents are already choosing open models when they pay per token, that's the demand side of the Bittensor hypothesis showing up in real numbers.
This chart shows the world moving in the direction Bittensor needs. That's why I say the data backs it.


The fair pushback on that data
Now, this isn't totally fair, as it only looks at API usage through OpenRouter.
So Google searches that use AI, chatting with ChatGPT or Claude, or using Grok on X is not taken into account.
It only looks at the paid usage by businesses and personal agents.
... but isn't that more important...?
I think it is.
Here's my reasoning. A person chatting with ChatGPT for free doesn't pay anyone. A business running an agent pays per token and picks the model that does the job for the lowest price. That's the kind of usage that decides where the money goes in the long run.
There are two more objections I take seriously though.
First, OpenRouter is one router. It's popular with developers and agent builders, and those people are more likely to try open models than a bank is. So the real share across all paid usage is probably lower than 75%. I still think the direction holds.
Second, price. A lot of this flip is open models being much cheaper per token. If the big guys cut prices hard, some of that usage comes back. If the share starts rolling back towards 50%, I'll take that seriously.
And one more. Even if open source wins, that doesn't mean Bittensor is the network that captures it. That's the risk I'm paying 15% of my portfolio for, and I'll get to it below.
TAO price: a 17X, a 75% drawdown, and now a 48% drop
Price-wise, it hasn't really done much since the 17X rally in late 2023 going into 2024.
The weekly chart is below.
The rally was 1,602% from the 2023 low to the early 2024 top. Since then it's down 75% from the high around $740.
And in the past couple of months it dropped another 48%, from around $340 down to where it is now.
As I write this, TAO is $186.76. The all-time high was $757.60, so the 75% drawdown still stands. It's down 31.6% in the past 30 days and 14.2% in the past 7 days.
It's also 22% below its 200-day average and 27% below its 50-day. So no, this isn't a chart that's turning yet.
What I like about the chart is the box. Since early 2025 TAO has made a lower high every time (the upper white line), but the lows have held between roughly $175 and $200. We're at the bottom of that box right now.
The lower white line is the one I care about. It slopes down from the 2024 low, and it sits a bit under current price. If TAO loses the box, that line is the next place it should find buyers, somewhere around $150 to $170 on the chart.
The market cap is around $2.2 billion. For a network that's trying to be the open alternative to the biggest industry on the planet, I think that's small.
Bittensor is risky, but I think this has the long-term potential to go 5X-10X at least over the next few years.
To put a number on that: 5X from here is around $930, which is above the 2024 high. 10X is close to $1,900. I don't expect that next year. I think it's where the network's value goes over a few years if the open source trend keeps running.

Where TAO goes wrong
I said risky. Let me be specific.
TAO already did one 75% drawdown. In a bear market, alts that have already fallen 75% can fall another 50% from there. The chart doesn't protect me from that.
The second risk is the one from above. Open source can win and Bittensor can still lose. The network needs the open models people actually pay for to be built and served through Bittensor. If that happens somewhere else, owning TAO doesn't pay off even though I'm right about the trend.
The third risk is that I'm early. Decentralized AI is a multi-year story. It can sit in this box for another year while BTC recovers, and 15% of a portfolio going nowhere has a cost.
What would change my view?
If the open weight share on OpenRouter rolls back under 50% and stays there, half my reason for owning it is gone.
If TAO breaks the lower trend line on weekly closes and keeps going, I'd have to accept the market is pricing in something I'm not seeing.
Neither has happened. So I'm buying.
You've just read the whole Bittensor case as it stands today: the open weight share flipping from 25% to 75% on OpenRouter, the fair objections to that data, the box on the weekly chart, and the two things that would make Oskar change his mind. This article is a snapshot, and it can't show what happens after today. The open source share moves, TAO either holds the box or loses it, and the thesis gets re-examined. Inside the Solberg Invest community, Oskar Solberg walks through that re-examination as it happens, in the same plain reasoning you've just read, so you can see how the thinking changes over time. See how the Solberg Invest community works
Bet two: Hyperliquid, and the move to tokenize everything
Hype is the proof that web3 project can break free from the bad reputation of crypto.
It's a derivatives exchange that runs on-chain, it's fast, and people actually use it. That alone puts it on a very short list.
The way I see it, the move to tokenize assets will take Hyperliquid to hundreds of dollars.
Here's the logic. Right now, Hyperliquid makes money on crypto derivatives. If stocks, commodities and other assets get tokenized and traded on-chain, the same exchange can list them. The volume available to it goes from crypto to everything.
And derivative markets are enormous compared to crypto.
Don't believe me?
Read up on the size of derivative markets.
Then check out the deals Hyperliquid has made with for example Intercontinental Exchange. Those deals tell you who they're building for. If the thesis is "traditional assets end up trading here", those are the partners you want.
So getting in at $60 is a bargain, in my view.
The numbers on HYPE
Let's look at price first. The weekly chart is below.
HYPE topped out just under $98 in September. This week it tapped $60, which is a drop of almost 40%. That's the box on the chart.
As I write this, it's back around $65.45. Market cap is roughly $15.6 billion.
Notice the oscillator at the bottom of the chart. It's still positive for HYPE (8.07), while it's deep negative for BTC and TAO. That shows HYPE has held up far better over the past year than the rest of the market. It went from around $20 in early 2026 to almost $100 while BTC was going sideways and down.
So 40% off is a big discount on a strong chart. It's much easier to justify than just a few short months ago.
Now the valuation.
Yes, the market cap is $15B. But they're burning almost a billion dollars worth of tokens annually.
Put those together. A billion a year against $15.6B is a burn of roughly 6% of the market cap every year, at today's price. That burn is the reason I'm fine paying $15B for it.
And I think the revenue (and burns (....and price!)) will go 5X-20X over the next 3-5 years.
If revenue goes 5X and the market keeps paying the same multiple, that's $300+ per token. That's how I get to "hundreds of dollars" from $65.

Where HYPE goes wrong
The risk with Hyperliquid is simpler than the risk with Bittensor, but it's bigger in the short term.
Revenue comes from trading volume. In a bear market, volume drops. So the burns I just used to justify the valuation can shrink exactly when the price is falling. That's a double hit.
Second, competition. Every big exchange and several chains want this business. Hyperliquid has the lead right now, and leads in crypto don't always last.
Third, the chart. HYPE bounced at $60 this week, but the next real support is the box from 2025. That's $40 to $50 on the chart. If this bear has one more leg, I expect HYPE to visit it.
That's also my next buy zone. I'm not waiting for it, because I don't know if we get there. But if we do, I'll be adding.
What would change my view?
If revenue and burns stop growing for a few quarters while competitors take share, the "hundreds of dollars" case is gone. At that point I'd treat it as a trade and size it like one.
The burn-and-revenue check is the kind of work Oskar repeats each quarter
The Hyperliquid case rests on a few numbers that will move: revenue, token burns, and whether the lead over competitors holds. Oskar has laid out the full logic for buying at today's level, the next zone where he'd add, and the exact condition that would turn this from a long-term holding into a trade. Those checks don't end with this article. In the Solberg Invest community, Oskar Solberg revisits the same thesis as new quarters of data arrive and explains whether the reasoning still holds, in the same macro and fundamental framing used here. That ongoing check-in is part of what the community offers. Learn about the Solberg Invest community
What I'm watching from here
To tie it together.
I'm buying two long-term bets in Q4 of the bear market because they're cheap, and because the thing each one needs is already happening.
For Bittensor, the thing is open source winning. That chart went from 25% to 75% in a year.
For Hyperliquid, the thing is on-chain derivatives and tokenized assets. The burns show it's already making real money, and the deals show where it's headed.
What I'll be checking:
The open weight share on OpenRouter. Above 65% means the trend holds. Rolling back towards 50% means I'm wrong about the direction.
Hyperliquid's revenue and burns, quarter by quarter. They need to grow through the bear too.
BTC holding the orange box. If BTC loses it with volume, every alt gets hit again, and I'll be glad I kept 20% in USDC.
HYPE at $40-$50 and TAO at the lower trend line. Those are where I add.
I'm not calling an exact bottom here. The cycle keeps on cycling. I'm buying now.
