←
Dhiraj15,639 posts
Dhiraj avatar

Dhiraj

@IndustrlPolicy

Interested in geopol & external sector. Schumpeterian capitalism | Environmentalism | Ambedkarite

◷ Joined January 2023▧ 1,406 Photos and videos
Bill Bishop avatar
⟳ Dhiraj retweeted
Bill Bishop@niubi·Nov 17

“China has offered additional tariff relief for about C$400mn ($280mn) in Canadian exports of frozen pork, fish and seafood, while Ottawa will extend tariff relief for two years on Chinese steel and aluminium products that are set to expire at the end of 2026.”

Xi and Canada’s Carney slash tariffs as China courts middle powers ft.trib.al/kvMDD76

Dhiraj avatar
Dhiraj@IndustrlPolicy·12h

plant is now operating following its stabilization phase and it is showing in sudden spike in trade is concentrated in HSN codes corresponding to memory chips.

Bloomberg avatar
⟳ Dhiraj retweeted
Bloomberg@business·Nov 17

Putin’s war against Ukraine is undermining the president’s own efforts to halt the country’s demographic decline. By 2025, the conflict had become the most common cause of death for working-age men born across most of Russia, according to confidential data held by Rosstat. Read more: bloom.bg/4t2ZQKi 📷: Sergei Bulkin/AFP

Post attachment
Martin Chorzempa 马永哲 avatar
⟳ Dhiraj retweeted
Martin Chorzempa 马永哲@ChorzempaMartin·Nov 17

I’ve been saying, CN AI models are departing in major ways from open source 1) some top models are not open weight (Ali is one ex, another is Seedance) 2) Kimi and others now charging cloud providers for hosting 3) export controls limiting open weight cyber, so firms tune down cyber capabilities before releasing weights

Alibaba's new Qwen4-Max model will not be open-source. They say: "Smaller models in the family are coming to open source, built on the same lineage and trained with the same generation of techniques, tuned for what small models do best."

Dhiraj avatar
Dhiraj@IndustrlPolicy·Nov 17

Indian pharma is pretty much following same route as they first consolidated their position in generics and now Wockhardt and Glenmark etc are moving into novel products. Jio will pretty much follow same pattern.

Dhiraj avatar
Dhiraj@IndustrlPolicy·Nov 17

In a nutshell, if a cos is steadily increasing its capital stock to create economies of scale then it is generally a good sign even if it not in R&D spending (Indian cos' R&D accounting is often unreliable anyway).

Dhiraj avatar
Dhiraj@IndustrlPolicy·Nov 17

path from lab to pilot is littered graveyard of promising projects that never make it to commercial scale. The one thing one can learn from is Chinese is first capture the market with scale then go do rest the work.

Dhiraj avatar
Dhiraj@IndustrlPolicy·Nov 17

As stationary grid storage, it is cost and reliability matter more than cutting edge chemistry and this LFP does not need to be Gen 3 or above as once grid storage market is consolidated then captive profits can be redeployed to wunderwaffe ventures.

Dhiraj avatar
Dhiraj@IndustrlPolicy·Nov 17

RIL acquired its sodium-ion battery IP through Faradion but it has struggled to scale to commercially required yields. It is eg of a frontier R&D product that has yet pay off because lacking scale. A better investment would be high yield generic LFP manuf to consolidate BESS mkt.

Beating scale is not easy. If a market becomes a perfect competition or a duopoly then doubling down on R&D is must. Eg Ind generic pharma has scale and no random upstart can simply walk in and challenge it because

Dhiraj avatar
Dhiraj@IndustrlPolicy·Nov 17

primary goal should be to capture the mkt and scale until it consolidates into a duopoly or perfect competition and only then double down on R&D to beat your apex rival. Until that investing capex in capital stock is a better bet than funding a uncertain project from lab to pilot

Dhiraj avatar
Dhiraj@IndustrlPolicy·Nov 17

Also in war your most R&D intensive produced wunderwaffe systems can be beaten by sheer volume as we have seen here too. Chinese primarily play at scale.

Dhiraj avatar
Dhiraj@IndustrlPolicy·Nov 17

that scale also has accumulated tacit know-how built up over decades and that knowledge isn't captured in filed IP or patents. You can invent a wunderwaffe battery chemistry but if you can't scale it with acceptable yield, you'll lose to a gen old tech that can.

Dhiraj avatar
Dhiraj@IndustrlPolicy·Nov 17

Beating scale is not easy. If a market becomes a perfect competition or a duopoly then doubling down on R&D is must. Eg Ind generic pharma has scale and no random upstart can simply walk in and challenge it because

Dhiraj, the problem with capex is that capacity can be bought by anyone with a balance sheet, so the returns on it get competed down over time. R&D is high variance, agreed, but it is the only route to pricing power. Take Indian IT for example, decades of low R&D and now most of them are price takers.