Uzbekistan's hard currency comes from four places. Two of them are cyclical.

The Central Bank of Uzbekistan buys around nine tenths of the gold mined in the country, sells it abroad, and uses the foreign currency it earns to mop up the soum it printed to pay for it.

That loop keeps the lights on. It is one of two pillars holding up the country's supply of foreign exchange. The other is the roughly $19bn a year sent home by Uzbeks working abroad.

On the central bank's own account, four things bring hard currency into Uzbekistan: gold, remittances, foreign direct investment, and money raised abroad from creditors, international financial institutions and bond markets. The first two are much the largest. They are also the two that answer to the world gold price and to the Russian labour market rather than to anything decided in Tashkent.

Both are now flagged as turning — which makes the two smaller pillars the most consequential thing about the next five years.

The gold loop

"The central bank, as the regulator and monetary authority, has a privileged right to buy that gold," the head of the bank's FX regulation and supervision department told Isaac Hanson of Perspective Publishing. "We buy around 90 to 92% of the overall amount."

Two state-owned miners, Almalyk and Navoi, produce around 100 tonnes a year between them. What the bank leaves goes to 25 or 30 domestic jewellery firms, whose processed exports have grown over the past 18 months.

The purchase is made in soum, and that is the part that matters. Buying domestic gold injects local currency into the economy, and left alone it becomes inflation. So the bank sells the gold abroad, brings the foreign currency home and sells it into the domestic market, which takes the soum back out. Its own bonds, placed domestically and closed to non-residents, do the same job in parallel. "At the end of the day we end up with a neutrality principle," he said.

The outline is not new: the IMF's Article IV documentation records the priority right, notes that the bank supplies the currency market in amounts equivalent to what it has bought, and states that its sales are aimed at sterilising liquidity rather than steering the rate. What has not been public is the granularity — and it arrives ahead of the intervention strategy Governor Timur Ishmetov has promised to publish.

There is a catch in it. The selling leg is discretionary and timed. Uzbekistan suspended gold exports for six months, resumed briefly in April and paused again in May. Exports for January to May came to $1.5bn against $6.49bn a year earlier, dragging total exports down 15.5%. The bank waits for prices rather than selling to a schedule, which means that for much of this year the absorption has had to be carried by bond issuance instead.

What sits in the vault

Reserves stood at $64.34bn on 1 August. Gold accounted for $56.12bn of that, or 87%.

The concentration has been building for a decade. Gold was under half of Uzbek reserves in 2016 and had reached nearly two thirds by 2024, when analysts were already uneasy about the reliance. There is no target for it, the FX official said: "there is no strategy to keep like 85 or 80% of gold in the reserve." On 1 August the figure was above both.

What that exposure means was demonstrated in June. The gold price fell from $4,516.85 to $4,016.70 a troy ounce and reserves dropped $6.8bn in a single month, from $70.6bn to $63.8bn. The bank put the negative revaluation effect at $6.9bn, partly offset by continued buying. A tenth of the country's external buffer went on a price move, without a migrant losing a job or a tonne of metal leaving the vault. July recovered $590mn of it on a 0.9% rise.

The physical position is still growing — 41 tonnes bought in the first half of the year, nine in June alone, third in the world that month behind Poland and China. But the buffer against every other shock is now a bet on a single price.

Repositioning has begun. The securities holding inside the reserves fell around 40% during July, from $2.86bn to $1.76bn, offset by a $1.31bn rise in cash and deposits. Days after the Tashkent forum closed, Ishmetov told Bloomberg the bank would convene Goldman Sachs, BlackRock and JPMorgan in mid-September on investment options for the reserves. It is also examining sukuk, following June's Islamic finance law: "we adopted the law on Islamic finance and the central bank is now learning the opportunity to invest into Islamic sukuk bonds and other instruments," the FX official said. "Step by step, gradually, not at one time."

The migrants

For thirty years, the second pillar had one address. Uzbek labour migration meant Russia, and so did the money coming back from it — a flow now worth almost $19bn a year, close to 13% of GDP and around a quarter of all the foreign currency entering the domestic market.

In the space of about four years, that has started to come apart. Covid closed the borders, the invasion of Ukraine made Russia a harder and more dangerous place to work, and the war economy that followed has squeezed the wages that drew people there. The central bank's own quarterly data now shows the Russian share falling five points in twelve months:

Source of remittances   Q1 2025     Q1 2026

Russia        77.6%        72.4%

Kazakhstan 3.1% 4.1%

South Korea        3.5% 4.1%

Europe        2.3% 3.3%

Other 13.6%        16.2%

Source: Central Bank of Uzbekistan labour market review, Q1 2026

The trend runs longer than the table. Russia accounted for 87% of the flow a few years ago and 78% by 2024, and the shift has been building steadily as Uzbekistan's labour ministry has worked with governments in Europe, East Asia and the Middle East to open alternatives. There are now 48 agreements with 23 countries on organised labour migration. One operator serving the Uzbek diaspora reported UK volumes up 67% and European volumes up 87%, with $416mn sent in the first half of this year.

For a country where this money is load-bearing — World Bank research finds that without it Uzbekistan's poverty rate would rise by three quarters, from 9.6% to 16.8% — spreading the risk across five destinations instead of one ought to be unambiguously good news. The total is rising too: $9.3bn in the first six months of 2026, against $6bn for the whole of 2020.

But the sharpest argument anyone made about it was that the shift may not hold, and nobody answered it.

Migration follows gravity, a development economist argued: people move to places that are geographically or culturally close, and remittance patterns follow. In most of the world diversification happens through a pull factor, opportunity at the destination drawing people in. What happened across Central Asia after 2022 was the opposite. It was a push. People left Russia because conditions pushed them out, not because Seoul or London pulled them in.

That distinction decides whether the infrastructure now being built is permanent or temporary. If the push reverses, gravity pulls the migrants back and the corridors, licences and integrations being constructed today are stranded assets. Holding them, he said, needs genuinely better jobs at the far end — convenience alone will not do it.

Kyrgyzstan shows the downside. The number of Kyrgyz migrants in Russia fell from 270,000 in November 2022 to 170,000 by November 2024, and what made it damaging was not the scale of the fall but that it was not compensated anywhere else. A hundred thousand people left the corridor and did not appear in another one.

Roman Rybalkin of S&P Global Ratings, on a different platform, reached the same place independently: labour markets in the countries where the region's remittance income originates were unlikely to stay tight indefinitely, and there were early signs of trouble in them. Analysts have warned for over a year that a cooling Russian war economy would be felt hardest in Uzbekistan, Tajikistan and Kyrgyzstan.

The two that could balance it

Against those, the smaller pillars look increasingly like the answer rather than the afterthought.

Foreign direct investment reached $8.3bn in the first quarter of this year. And Uzbekistan has spent seven years learning to borrow: the sovereign made its eurobond debut in 2019 with $1bn in two tranches, and the state banks followed. "Most of the state-owned banks already attracted," the FX official said. "We have in portfolio SQB, the National Bank of Uzbekistan. Recently Alokabank was a pioneer — this year they issued successfully." The gold miners themselves issue eurobonds and are listed in London. Earlier this year the sovereign printed a local-currency eurobond worth roughly $1bn, the largest such deal across CEEMEA in fifteen years, and in May the National Investment Fund raised $603.6mn on the London and Tashkent exchanges, rising to around $691mn once the over-allotment was exercised.

Both pillars are young and neither is close to the scale of gold or remittances. Both, unlike gold and remittances, respond to what Uzbekistan does rather than to what happens to a commodity price or a war economy. That is the case for growing them fast.

Managing the volatility

The central bank's answer, in the meantime, is the exchange rate itself.

"Both appreciation and depreciation are harmful for the economy," the FX official said. "We use the FX rate as a shock absorber." The plan is not to influence it: "we put it as it is and let the market decide what is the optimal rate."

He has the anecdote to go with it. For years importers complained about annual depreciation of 9-12%. Last year the soum appreciated by around 7%, the first appreciation since independence, "and last year the exporters came to the central bank saying: now, guys, why is it appreciating so much? We have problems." Since daily volatility widened, he added, the complaints have largely stopped, because participants have started hedging — which he expects to force the development of a derivatives market.

But the shock absorber is not free. Samigjon Inogamov, director of the bank's monetary policy department, put a number on it to bne IntelliNews: the direct pass-through from the exchange rate to inflation is around 0.3, so a 1% depreciation adds roughly 0.3 percentage points to prices, rising to 0.4 or 0.45 with second-round effects.

A currency doing its job as a shock absorber is therefore also, mechanically, an inflation generator — in a country whose central bank has spent six years trying to break a public expectation of permanent devaluation. Historic double-digit inflation was driven substantially by continuous depreciation, Inogamov said, which is why household inflation expectations remain tied to devaluation expectations. The bank is holding the policy rate at 14% against inflation of 6.4% at the end of July, and expects to reach its 5% target next year for the first time since inflation targeting began in 2020.

The constraint written into law

One limit on the remittance pillar is entirely Uzbekistan's own. Under the law on currency regulation, all foreign exchange conversion must go through banks — which makes every non-bank money transfer operator a bank's customer, adds cost, slows settlement and deters new entrants in precisely the corridors the country most needs to grow.

A World Bank economist put the recommendation directly to the central bank: let money transfer operators manage their liquidity end to end. The answer was that only banks may currently convert currency, and that any easing would come later, within the capital account liberalisation roadmap being drafted with the IMF.

Cost is not the binding constraint on formalisation in any case. Sending money into this corridor already costs around 1%, against a 5% global average. What deters people is certainty. One operator described a transfer tripped by a financial crime check and held for questioning while a family waits at the other end; the sender is in a farm dormitory with forty or fifty other seasonal workers, and by morning all of them have switched to the man among them who will take the cash and guarantee it arrives.

The bank's own ambition for the flow goes further than fixing the plumbing. It is working with government agencies to turn remittances from money that funds consumption into money that funds savings and investment — and the delivery mechanism already exists in outline. Since last year commercial banks have posted staff in Uzbekistan's mahallas, the smallest administrative unit and roughly 9,000 in number, to work with residents on what products exist, how to calculate an effective interest rate and how to build a business plan. Whether that turns a transfer into a deposit is untested. The bank's chief financial inclusion officer, Dilbar Abduganieva, has warned that formal savings sit below both the regional and global benchmark, in a country she describes as culturally excellent savers.

Uzbekistan's record is strong: GDP up 8.5% in the first half of 2026 — fast enough that the central bank has warned of overheating and held rates rather than cutting — sovereign upgrades from S&P last year and Moody's in June, an exchange rate the IMF now classifies as floating, and state ownership of the banking sector down from 85% to close to 60%.

The reforms are real. The two things paying for them are not, on the evidence of the country's own ratings agency and its own central bank, permanent.